The primary objective of the proposed research is to test whether a novel technique, image-guided bone marrow-sparing intensity modulated radiation therapy (IG-BMS-IMRT), can reduce hematologic toxicity (HT) in patients receiving concurrent chemotherapy and pelvic radiation therapy (CRT). Radiation-induced HT is a significant clinical problem limiting the intensity of chemotherapy that can be delivered in patients with pelvic malignancies. Secondary objectives are to determine the functional properties of bone marrow (BM) subregions and to validate a novel MRI technique to quantify BM fat content. The specific aims are: (SA1) to test whether IG-BMS-IMRT will reduce HT for cervical cancer patients undergoing chemoradiotherapy; (SA2) to determine the functional properties of critical BM subregions in which models have found that increased radiation dose leads to increased HT; and (SA3) to validate the technical and biochemical properties of MR fat quantification techniques in bone marrow specimens. Our main hypothesis is that IG-BMS-IMRT will lead to a relative reduction of acute HT of 50% compared to historical controls. We will test this hypothesis in a prospective phase II multi-institutional clinical trial of 50 patients, with the primary endpoint of acute HT. We will acquire baseline 18F-3'-fluoro-3'-deoxy-L-thymidine positron emission tomography (FLT-PET) and quantitative fat fraction MRI scans to identify active BM subregions. We will segment the BM based on a previously developed protocol, then use this as a primary avoidance structure during IMRT planning. Dosimetric limits on active BM are derived from previous toxicity models we developed. We will obtain mid- and post-treatment fat fraction MRI scans to quantify changes in BM subregions. As a secondary analysis, we will test the hypothesis that FLT tracer uptake and changes in FF are greater in critical compared to non-critical BM subregions, using previously developed techniques based on high-dimensional data analysis and deformable image registration. For aim 2, we will test whether MR fat fraction estimates are stable under varying protocol conditions, and whether they correlate with biochemical and histological assays of fat content. The end result of this project will determine whether IG-BMS-IMRT has a strong physiologic rationale and clinically significant impact on toxicity, and will determine whether this approach should move forward to a phase III trial versus conventional RT.